You are on page 1of 5

MATEC Web of Conferences 224, 01108 (2018) https://doi.org/10.

1051/matecconf/201822401108
ICMTMTE 2018

Determination of shrinkage of weld


Nikolay Velikanov1, *, Sergey Koryagin1, and Oleg Sharkov1
1Immanuel Kant Baltic Federal University, 236041 Kaliningrad, Russia

Abstract. A calculation-experimental technique of determination of weld


shrinkage is described in that work. Experimentally, with the help of strain
gauges, residual stresses are determined at two points, which are at known
distances from the weld. Shrinkages are found out from the theoretical
solution of the plane elasticity theory about the insertion of bodies with
interference. Calculations have been carried out for experimental studies of
a weld 400 mm long for different thicknesses of metal. It is found that
when the thickness of the metal increases, the shrinkage values also
increase. This is due to the fact that as the thickness of the metal increases,
the heat input required to penetrate it also increases.

1 Introduction
Determining the shrinkage of welds is an urgent task. It is solved theoretically and
experimentally in a number of papers [1–16].
The aim of the paper [1] was to study the influence of various arc welding processes on
the development of microstructure, mechanical properties, residual stresses and
deformations in welded joints of sheets of austenitic stainless steel 9 mm thick. Welded
joints were made in 3 different modes of arc welding, which are used in the nuclear
industry. All the welds underwent X-ray examination. The microstructural characteristic
was determined using an optical and scanning electron microscope.
As shown in [2], welding deformations not only worsen the accuracy of the production
of hull blocks, but also reduce productivity due to correctional work. An accurate prediction
of the deformation caused by welding will help to control the accuracy of the structure
dimensions.
The work [3] shows that linear shrinkage of two-component specimens was greater than
the shrinkage of one-component specimens because the incompatibility of shrinkage of
both materials causes biaxial stresses and enhances bonding processes.
In [4], the process of joining of steel plates 25 mm thick with a narrow gap between
them. The joint was made using gas arc welding with the preservation of vertical building
up. The assembly of thick plates was also performed by the widely used multistage
multipass building up procedure. Shrinkages in welded joints were determined, and the
influence of the heat parameters of welding on shrinkage and bending was studied.
In some cases, the welded elements undergo stresses even before welding begins. The
initial stress state of structure, which can be caused by previous technological operations

*
Corresponding author: nvelikanov@kantiana.ru

© The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons
Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/).
MATEC Web of Conferences 224, 01108 (2018) https://doi.org/10.1051/matecconf/201822401108
ICMTMTE 2018

(including previous welding) or external loads, influence on the kinetics of deformations


and stresses and consequently on the shrinkage value. However, as experimental and
theoretical studies have shown [5–16], this influence can be neglected when calculating
stress fields in full-scale structures.

2 Calculation-experimental procedure

2.1 Accepted hypotheses and assumptions


Thus, the problem of determining residual stress fields in full-scale structure can be solved
by a superposition of the problem of determining the stresses from individual components.
These components include: external loads acting on the structure; prestresses before
welding; welding stresses; temperature difference of the structure individual parts, etc.

2.2 Theoretical model

The weld (Figure 1) is modeled by a rectangular insert in a plate with longitudinal and
transverse interferences equal to, respectively 2∆1 and 2∆ 2 (Figures 2, 3). Residual
stresses after welding are determined using the Kolosov-Muskhelishvili potentials [4]:

G g (t ) 
ϕ( z ) = ∫ dt , 
πi (χ + 1) L t − z 
 (1)
G ′
g (t ) + t g (t ) 
ψ( z ) = − ∫ t−z d t .
πi (χ + 1) L 

where z – is the complex coordinate of the point of the plate, z = x + iy ; G – is the


shear modulus of elasticity; χ = (3 − µ) (1 + µ) – is the coefficient expressed through the
Poisson constant µ ; t – is the point coordinate at the edge of the rectangular insert; g (t ) –
displacements at the edge of the insertion L .

Fig. 1. Weld scheme: 1 – strain gauges; 2 – weld; 3 – plate.

Expressions for displacements at the edge of insert g (t ) depend on the longitudinal and
transverse shrinkage of the insert: for the upper edge g (t ) = ∆1 x l + i∆ 2 , for the bottom
edge – g (t ) = ∆1 x l − i∆ 2 .

2
MATEC Web of Conferences 224, 01108 (2018) https://doi.org/10.1051/matecconf/201822401108
ICMTMTE 2018

Fig. 2. Longitudinal displacements at the edge of insert.

Fig. 3. Transverse displacements at the edge of insert.


For the accepted displacements at the insert edge, we obtain the following values of the
potentials from (1):

2G (∆1h − ∆ 2l ) l − z 
ϕ( z ) = ln , 
πi (χ + 1) l+z 
 (2)
− 2G  l − z 2lz (∆1h − ∆ 2l )  
ψ( z ) =  (3∆1h + ∆ 2l ) ln +  .
πi (χ + 1)  l+z z2 − l2  

Residual stresses are expressed through the potentials (2) by the Kolosov-
Muskhelishvili formulas.

2.3 The calculation algorithm

Weld shrinkage ∆1 and ∆ 2 are determined by the calculation-experimental method. After


weld deposit on the plate and its cooling, the strain gauges are attached to the plate surface
(Figure 1), their readings are taken.
Then, the sections of the plate are cut out together with the strain gauges and their
readings are taken again. By the difference in readings before and after the cutting, the
residual welding stresses are determined. According to these stresses, the weld shrinkages
are evaluated from the Kolosov-Muskhelishvili equations.

3 Results of calculations and discussion


From the data of the experiments, received by different authors, the values of normal
stresses G X at points with different coordinates and weld length 2l =400 mm for different
thicknesses of metal s were taken.

3
MATEC Web of Conferences 224, 01108 (2018) https://doi.org/10.1051/matecconf/201822401108
ICMTMTE 2018

Then these values were used to calculate the values of the longitudinal ∆1 and
transverse ∆ 2 shrinkages during the weld deposit and the welding of the stiffening rib.
In the calculations we took the value of the coefficient of transverse deformation
(Poisson's ratio) µ =0.3, the modulus of elasticity (the Young's modulus) E =2.1 105 MPa.
The shrinkage values obtained this way, depending on the metal thickness, are shown in
Figures 4, 5.
As follows from the figures, as the thickness of the metal increases, the shrinkage values
also increase. This is due to the fact that as the thickness of the metal increases, the heat
input required to penetrate it also increases, i.e. the electric current value in arc welding and
the gas rate in gas welding increase. The same factors explain the increase in shrinkage,
compared with manual welding, in case of semi-automatic and automatic welding.

Fig. 4. The shrinkage values during the welding of the seam: 1 – ∆ 2 ; 2 – ∆1 .

We also note that when the length of the weld increases, the transverse shrinkage
remains practically unchanged, and the longitudinal shrinkage increases in linear
dependence with the weld length.

Fig. 5. The shrinkage values during the welding of the rib: 1 – ∆1 ; 2 – ∆ 2 .

4
MATEC Web of Conferences 224, 01108 (2018) https://doi.org/10.1051/matecconf/201822401108
ICMTMTE 2018

The obtained graphical dependences of weld shrinkages on the metal thickness (Figures
4, 5) can be used to develop techniques aimed at preventing the destruction of hull
structures.

4 Conclusion
The above method of determination of residual welding stresses can be used in structure
individual parts while the thermal heating, warming up with a laser beam, etc.
The proposed methodology can be used not only to calculate the stress field in the
current construction, but also to create residual stresses fields of specified type in the
structure. This fact can be used to reduce stress concentrations in structure individual parts
to reduce post-repair stresses and in particular cases, for example, to remove camber,
embossing.

References
1. P. Vasantharaja, M. Vasudevan, V. Maduraimuthu, Transactions of the Indian institute
of metals, 71(1) ,127-137 (2018)
2. W. Liang, D. Deng, Advances in engineering software, 115, 439-451, (2018)
3. A. Manonukul, S. Songkuea, P. Moonchaleanporn, M. Tange, International journal of
minerals metallurgy and materials, 24(12), 1384–1393 (2017)
4. N.L. Velikanov, S.I. Koryagin, O.V. Sharkov, IOP Conference Series: Materials
Science and Engineering, 124, 012094, (2016)
5. B.P. Agrawal, P.K. Ghosh, Journal of the Brazilian society of mechanical sciences and
engineering, 39(12), 5205–5218 (2017)
6. M. Junaid, F.N. Khan, K. Rahman, M.N. Baig, Optics and laser technology, 97, 405–
419(2017)
7. H. Huang, S. Tsutsumi, J.D.Wang, L.Q. Li, Finite elements in analysis and design,
135, 1–10 (2017)
8. A.N. Vasileiou, M.C. Smith, J. Balakrishnan, J.A. Francis, C.J. Hamelin, Nuclear
engineering and design, 323, 309–316 (2017)
9. S. Seo, E.W. Huang, W. Woo, S.Y. Lee, International journal of fatigue, 104, 408–415
(2017)
10. S. Hossain, G. Zheng, C.E. Truman, D.J. Smith, Experimental techniques, 41(5), 483–
503 (2017)
11. S. Kainuma, M.Y. Yang, Y.S. Jeong, S. Inokuchi, A. Kawabata, D. Uchida,
Engineering failure analysis, 79, 520–537 (2017)
12. H.C. Yildirim International journal of fatigue, 101(2), 408–420 (2017)
13. A.I. Pritykin, A.S. Lavrova, Mechanika, 23(4), 488–494 (2017)
14. P. Yukhymets, A. Gopkalo, G. Zecheru, M. Mihovski, International journal of offshore
and polar engineering, 26(3), 272–277 (2016)
15. T.J. Kim, B.S. Jang, S.W. Kang, International Journal of Naval Architecture and Ocean
Engineering, 7(1) 157–173, (2015)
16. Y. Zhu, Y.S. Cheng, J. Liu, Chuan Bo Li Xue/Journal of Ship Mechanics, 18(10),
1228–1236. (2014)

You might also like